Multi-hop Question Answering on HotPotQA (Task Score, Brier)
57.6Task ScoreGEPA-C
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GEPA-COptimizer LLM=GPT-52026.05 | 57.6 | 0.41 | |
| GEPA-COptimizer LLM=GPT-5-mini2026.05 | 57.6 | 0.41 | |
| GEPA-COptimizer LLM=Gemini-3.1-Pro2026.05 | 57.6 | 0.41 | |
| GEPA-COptimizer LLM=Gemini-3.1-FL2026.05 | 57.6 | 0.41 | |
| RPTOptimizer LLM=GPT-52026.05 | 55.5 | 0.438 | |
| RPTOptimizer LLM=GPT-5-mini2026.05 | 55.5 | 0.438 | |
| RPTOptimizer LLM=Gemini-3.1-Pro2026.05 | 55.5 | 0.438 | |
| RPTOptimizer LLM=Gemini-3.1-FL2026.05 | 55.5 | 0.438 |